pub fn display_payloads_handles_native(
handles: &[u32],
lod_factor: f64,
) -> Vec<Result<DisplaySolidPayload, String>>Expand description
display_payload_handle_native for many resident solids at once, in
handles order.
Instances are meshed once. Solids that are rigid copies of one another
(an imported assembly places each part many times: 857 instances of 174
parts) share one tessellation: the copy with the finest display chord is
meshed, and each other copy gets that mesh and its edge polylines carried
by the rigid map between them, with its OWN face/edge ids and names and its
own vertices, so picking, selection and per-instance colour see the solid
they always saw. A copy is only drawn this way when everything but point
positions is identical (ids, senses, knots, weights, pcurves) and one rigid
map carries every vertex and control point of the meshed copy onto its own
(congruent_pose); anything else, and every solid with no copy, gets the
single call’s payload exactly.
Under the parallel feature the meshed solids are copied out of this
thread’s registry (it is thread-local) a chunk at a time and meshed across
rayon’s pool. Display meshing opens no thread-local mode (the soundness
scan’s mode is closed on every worker), so a payload does not depend on
the thread that computed it.